In MIMO multiplexing transmission, by transmitting different channel-coded data streams from a plurality of transmitting antennas using the same bandwidth, the data rate (i.e., the throughput) can be increased in proportion to the number of transmitting antennas within the given bandwidth. In S-PARC (Selective Per-Antenna Rate Control) employed in such an MIMO system, in addition to controlling the data rate adaptively in accordance with the change in radio wave propagation conditions, control is performed to maintain reception quality by stopping signal transmission from any transmitting antenna for which the received SINR (Signal-to-Interference and Noise power Ratio) has dropped below a predetermined value and by concentrating the transmitter power to the remaining transmitting antennas.
In this case, when the number of antennas used for transmitting the data decrease, the number of channel estimates needed for computation at the receiving end decrease, and as a result, the amount of computation decreases. In other words, the computing resources allocated to the path no longer in use are rendered idle. On the other hand, the channel estimates are periodically calculated from the received data and used for the computation for channel demultiplexing. For example, as described in JP 2001-522166A, if the calculated channel estimates are linearly interpolated to reduce the range of received data to which the same channel estimate is applied, it is possible to quickly cope with fast fading, but this requires a large amount of computing resources.